In this experimental research on persuasive communication and drug abuse prevention, we will generate and test theory-based message strategies for prevention campaigns within the context of theories of attention and persuasion. Messages will be designed to meet needs for novelty and sensation, which influence initial alerting and attending processes, and needs for cognition, which influence motivation for higher level processing of the information. Specific aims of the research are to investigate the extent to which message sensation value and message cognition value influence message processing and attitude, behavioral intention and behavior change in relation to marijuana; to assess possible interactions between types of messages and individual differences in needs for cognition and needs for novelty and sensation, and to assess age-related differences in message processing and response. Extensive formative research will be conducted to assess indicators for message content with high versus low cognition value, and to create, assess, and produce messages. Messages will be designed to meed cognition value criteria, as well as for high vs low sensation value. Three experiments are proposed to evaluate message effects. The first experiment will be conducted in the Communication Laboratory with 360 randomly selected college students and non-college young adults, who will view four videotaped anti-marijuana messages comprising one of four strategies (high vs low message sensation value by high vs low cognition value) 4 times each (16 exposures) over an 8-week period. Cognitive processing will be assessed through self- reports and psychophysiological measures. Attitudes and behavioral intentions will be assessed through self-report, and behavior will be assessed via self-report and urinalysis. The second experiment with 40 additional young adults will be conducted in the Communication Laboratory, where messages will be shown, and in the Residential Research Facility where marijuana use will be observed and verbal reports of drug effect, task performance, and cardiovascular activity will be taken. The third experiment will follow the same procedures used in the first with 240 middle school students, except that adapted messages will be viewed at school sites and adapted measures will be used.